Robust optimization for adversarial learning with finite sample complexity guarantees
Decision making and learning in the presence of uncertainty has attracted significant attention in view of the increasing need to achieve robust and reliable operations. In the case where uncertainty stems from the presence of adversarial attacks this need is becoming more prominent. In this paper w...
Main Authors: | , , |
---|---|
Format: | Conference item |
Language: | English |
Published: |
IEEE
2024
|
_version_ | 1811140272359211008 |
---|---|
author | Bertolace, A Gatsis, K Margellos, K |
author_facet | Bertolace, A Gatsis, K Margellos, K |
author_sort | Bertolace, A |
collection | OXFORD |
description | Decision making and learning in the presence of
uncertainty has attracted significant attention in view of the
increasing need to achieve robust and reliable operations.
In the case where uncertainty stems from the presence of
adversarial attacks this need is becoming more prominent.
In this paper we focus on linear and nonlinear classification
problems and propose a novel adversarial training method
for robust classifiers, inspired by Support Vector Machine
(SVM) margins. We view robustness under a data driven lens,
and derive finite sample complexity bounds for both linear
and non-linear classifiers in binary and multi-class scenarios.
Notably, our bounds match natural classifiers’ complexity. Our
algorithm minimizes a worst-case surrogate loss using Linear
Programming (LP) and Second Order Cone Programming
(SOCP) for linear and non-linear models. Numerical experiments
on the benchmark MNIST and CIFAR10 datasets show our
approach’s comparable performance to state-of-the-art methods,
without needing adversarial examples during training. Our work
offers a comprehensive framework for enhancing binary linear
and non-linear classifier robustness, embedding robustness in
learning under the presence of adversaries. |
first_indexed | 2024-09-25T04:19:21Z |
format | Conference item |
id | oxford-uuid:b4576446-cc3e-417e-a79c-4a53bb25279b |
institution | University of Oxford |
language | English |
last_indexed | 2024-09-25T04:19:21Z |
publishDate | 2024 |
publisher | IEEE |
record_format | dspace |
spelling | oxford-uuid:b4576446-cc3e-417e-a79c-4a53bb25279b2024-07-25T14:59:37ZRobust optimization for adversarial learning with finite sample complexity guaranteesConference itemhttp://purl.org/coar/resource_type/c_5794uuid:b4576446-cc3e-417e-a79c-4a53bb25279bEnglishSymplectic ElementsIEEE2024Bertolace, AGatsis, KMargellos, KDecision making and learning in the presence of uncertainty has attracted significant attention in view of the increasing need to achieve robust and reliable operations. In the case where uncertainty stems from the presence of adversarial attacks this need is becoming more prominent. In this paper we focus on linear and nonlinear classification problems and propose a novel adversarial training method for robust classifiers, inspired by Support Vector Machine (SVM) margins. We view robustness under a data driven lens, and derive finite sample complexity bounds for both linear and non-linear classifiers in binary and multi-class scenarios. Notably, our bounds match natural classifiers’ complexity. Our algorithm minimizes a worst-case surrogate loss using Linear Programming (LP) and Second Order Cone Programming (SOCP) for linear and non-linear models. Numerical experiments on the benchmark MNIST and CIFAR10 datasets show our approach’s comparable performance to state-of-the-art methods, without needing adversarial examples during training. Our work offers a comprehensive framework for enhancing binary linear and non-linear classifier robustness, embedding robustness in learning under the presence of adversaries. |
spellingShingle | Bertolace, A Gatsis, K Margellos, K Robust optimization for adversarial learning with finite sample complexity guarantees |
title | Robust optimization for adversarial learning with finite sample complexity guarantees |
title_full | Robust optimization for adversarial learning with finite sample complexity guarantees |
title_fullStr | Robust optimization for adversarial learning with finite sample complexity guarantees |
title_full_unstemmed | Robust optimization for adversarial learning with finite sample complexity guarantees |
title_short | Robust optimization for adversarial learning with finite sample complexity guarantees |
title_sort | robust optimization for adversarial learning with finite sample complexity guarantees |
work_keys_str_mv | AT bertolacea robustoptimizationforadversariallearningwithfinitesamplecomplexityguarantees AT gatsisk robustoptimizationforadversariallearningwithfinitesamplecomplexityguarantees AT margellosk robustoptimizationforadversariallearningwithfinitesamplecomplexityguarantees |